Identification of Ser153 in ICL2 of the Gonadotropin-releasing Hormone (GnRH) Receptor as a Phosphorylation-independent Site for Inhibition of Gq Coupling*

Type I gonadotropin-releasing hormone (GnRH) receptor (GnRHR) is unique among mammalian G-protein-coupled receptors (GPCRs) in lacking a C-terminal tail, which is involved in desensitization in GPCRs. Therefore, we searched for inhibitory sites in the intracellular loops (ICLs) of the GnRHR. Synthetic peptides corresponding to the three ICLs were inserted into permeabilized αT3-1 gonadotrope cells, and GnRH-induced inositol phosphate (InsP) formation was determined. GnRH-induced InsP production was potentiated by ICL2 > ICL3 but not by the ICL1 peptides, suggesting they are acting as decoy peptides. We examined the effects of six peptides in which only one of the Ser or Thr residues was substituted with Ala or Glu. Only substitution of Ser153 with Ala or Glu ablated the potentiating effect upon GnRH-induced InsP elevation. ERK activation was enhanced, and the rate of GnRH-induced InsP formation was about 6.5-fold higher in the first 10 min in COS-1 cells that were transfected with mutants of the GnRHR in which the ICL2 Ser/Thr residues (Ser151, Ser153, and Thr142) or only Ser153 was mutated to Ala as compared with the wild type GnRHR. The data indicate that ICL2 harbors an inhibitory domain, such that exogenous ICL2 peptide serves as a decoy for the inhibitory site (Ser153) of the GnRHR, thus enabling further activation. GnRH does not induce receptor phosphorylation in αT3-1 cells. Because the phosphomimetic ICL2-S153E peptide did not mimic the stimulatory effect of the ICL2 peptide, the inhibitory effect of Ser153 operates through a phosphorylation-independent mechanism.

The desensitization mechanism of GPCRs 1 involves binding of the agonist to the receptor, which not only causes activation of the effector but also facilitates phosphorylation of the receptor by second messengers-activated kinases or by G-proteincoupled receptor kinases (GRKs) on specific sites within the C-terminal tail. The phosphorylation enables binding of ␤-arrestin, which prevents further effector activation and targets the desensitized receptor for internalization via clathrin-coated vesicles that are pinched off from the plasma membrane by dynamin (1,2). The receptor is then dephosphorylated by a GPCR phosphatase (3) and can be targeted to lysosomes for degradation or recycled to the cell surface (1,2).
A striking feature of mammalian type I GnRHR is the absence of a C-terminal tail (4). Therefore, it is thought that the GnRHR does not undergo C-terminal tail phosphorylation, rapid desensitization, and internalization (for review see Ref. 5). Therefore, the known desensitization of gonadotropin secretion observed during sustained GnRH administration is attributed to down-regulation of inositol 1,4,5-trisphosphate receptors, desensitization of Ca 2ϩ mobilization, reduction in the number of GnRHR, and G q/11 expression and attenuation of phospholipase D activation and arachidonic acid release (5). In addition, prolonged GnRH treatment down-regulates PKC, cAMP, and Ca 2ϩ -dependent signaling (6). Because the GnRHR lacks a C-terminal tail, we reasoned that other structural components, such as the ICLs, might be involved in signal termination of the GnRHR as outlined above.
Synthetic peptides are important tools for understanding the sites and mechanisms of receptor/G-protein interactions. It was shown that peptides derived from the ICL2, ICL3, and ICL4 loops of rhodopsin disrupt the ability of G t to stabilize the active signaling conformation of rhodopsin and metarhodopsin II (7,8). In another study it was shown that a dimer of peptides from the C-terminal and N-terminal regions of ICL3 of the ␣ 2A -adrenergic receptor affect the stimulation of the guaninenucleotide exchange protein of G o (9). Although a motif scan of the human GnRHR identifies several potential phosphorylation sites in the intracellular loops, including sites for protein kinase A (Ser 74 , Thr 84 , and Thr 265 ), PKC (Thr 40 , Thr 42 , Thr 51 , Ser 74 , Thr 84 , Ser 118 , Thr 265 , and Thr 274 ), and calmodulin-dependent kinase 2 (Thr 265 ), it is not known whether these sites are functional and involved in receptor regulation (4).
We have therefore utilized synthetic peptides corresponding to the ICLs of the mammalian type I GnRHR in order to identify potential inhibitory sites involved in GnRHR function. Here we report that Ser 153 is a key residue in an inhibitory domain in ICL2, which exerts its inhibitory effect in a phosphorylation-independent manner.

EXPERIMENTAL PROCEDURES
Materials-The stable GnRH agonist Buserelin was used throughout this study and was kindly provided by Dr. J. Sandow (Aventis Pharma, Hoechst, Frankfurt, Germany). Dulbecco's modified Eagle's medium (DMEM), M199 medium, horse serum, fetal calf serum (FCS), penicillin/streptomycin antibiotics, trypsin, EDTA, and trypan blue were purchased from Biological Industries (Kibbutz Beit Haemek, Israel). OPTI-FLUOR scintillation liquid was purchased from Packard Instrument Co. Oligonucleotide primers for site-directed mutagenesis were purchased from Eisenberg Bros. (Israel). The Wizard Plus Mini-and Midipreps DNA purification kits were purchased from Promega. Jet-Sorb DNA kits for extraction from agarose gels were purchased from Genomed (Germany). Restriction-and DNA-modifying enzymes were from New England Biolabs. The cDNA of the influenza hemagglutinin (HA) epitope-tagged rat GnRHR in pcDNA3 plasmid was kindly provided by Dr. J. Neill (University of Alabama). Mouse monoclonal anti-HA epitope tag, secondary antibody goat anti-mouse, mouse monoclonal anti-active (doubly phosphorylated) ERK, and polyclonal antibodies to general ERK were from Sigma. Secondary antibody goat anti-rabbit was purchased from Jackson ImmunoResearch. myo-[ 3 H]-Inositol (80 -100 Ci/mmol) was purchased from Amersham Biosciences.
[ 125 I-D-Trp 6 ]GnRH (GnRH-A) was kindly provided by Dr. Y. Koch (The Weizmann Institute, Rehovot, Israel). Nitrocellulose sheets were purchased from Schleicher & Schuell. Immunoblots reagents were purchased from Bio-Rad. Other reagents were of analytical grades and were purchased from Sigma or Merck.
Peptide Synthesis-Peptides were synthesized on an ABIMED AMS 422 multiple peptide synthesizer (Langenfeld, Germany), employing the N-(9-fluorenyl)methoxycarbonyl (Fmoc) strategy following the commercial protocols. Peptide chains assembly was conducted on a 2-chlorotrityl chloride resin (Novabiochem). Crude peptides were purified to homogeneity by reverse-phase high pressure liquid chromatography on a semi-preparative silica C-18 column (250 ϫ 10 mm; Lichrosorb RP-18, Merck). Elution was accomplished by a linear gradient established between 0.1% trifluoroacetic acid in water and 0.1% trifluoroacetic acid in 70% acetonitrile in water (v/v). The compositions of the products were determined by amino acid analysis (Dionex automatic amino acid analyzer, Sunnyvale, CA) following exhaustive acid hydrolysis. Molecular weights were ascertained by mass spectrometry (VG Tofspec; Laser Desorption Mass Spectrometry; Fison Instruments, Manchester, UK).
Construction of GnRH Receptor Mutants by PCR-based Site-directed Mutagenesis-A 1.038-kb rat GnRHR cDNA tagged at the N terminus with the influenza HA epitope inserted at the KpnI and XhoI sites of pcDNA3 served as a template for creating site-directed mutations (10). The mutations were performed using separate primers. The plasmid template DNA (ϳ0.5 pmol) was added to a PCR mixture containing, in 25 l of 1ϫ mutagenesis buffer (20 mM Tris-HCl (pH 7.5), 8 mM MgCl 2 , 40 g/ml bovine serum albumin), 12-20 pmol of each primer, 250 M each dNTP, 5 units of Pfu Extender (Stratagene). The PCR cycling parameters were 1 cycle for 5 min at 94°C followed by 25 cycles of 20 s at 94°C, 30 s at 53°C, and 1 min at 72°C, followed by a final step of 10 min at 72°C. The PCR product was loaded into 1% agarose gel, and the GnRHR cDNA band (1.03 kb) was cut and cleaned using "JetSorb" DNA purification kit (Promega). Mutant cDNAs were subcloned into the KpnI/XhoI sites of the mammalian expression vector pcDNA3/Amp (Invitrogen). Products of the mutagenesis were used to transform competent XL-1 Blue Escherichia coli. Plasmid DNA was extracted from ampicillin-resistant clones and sequenced.
Transfection of COS-1 Cells-Plasmid DNA for transfection was prepared using Promega columns according to the manufacturer's instructions. COS-1 cells were cultured in DMEM (Invitrogen) containing 10% FCS in a 10% CO 2 atmosphere. Cells were seeded at 2 ϫ 10 6 cells per well in a poly-D-lysine-coated 20-cm plate 1 day before transfection by a modified DEAE-dextran method. Cells were washed twice with PBS and then incubated with 2 ml/well PBS containing 20 g/ml plasmid DNA and 0.2 mg/ml DEAE-dextran for 30 min at 37°C. The cells were incubated for a further 2.5 h at 37°C with DMEM containing 10% fetal calf serum and 100 mmol/liter chloroquine, after which they were washed twice with PBS containing 10% Me 2 SO and twice with PBS. The cells were then cultured overnight in DMEM with 10% fetal calf serum. Cells were then harvested and seeded in 6-well plates.
Phosphoinositide Hydrolysis Assay-The cells (5 ϫ 10 6 /well) were cultured for 2 days in DMEM, 10% FCS, antibiotics, and myo-[2-3 H]inositol (1.5 Ci/ml). The cells were then washed three times with DMEM containing 0.1% bovine serum albumin and incubated for 15 min at 37°C with 1 ml of the same buffer containing 10 mM LiCl. The cells were then treated with the corresponding concentration of GnRH. Reactions were stopped by aspiration of the medium and addition of 0.25 ml of H 2 O. Cells were scraped and transformed to tubes containing 1 ml of chloroform/methanol (1:2). Following incubation for 30 min at room temperature, 350 l of chloroform and 350 l of water were added, and the cells were centrifuged for phase partition (2500 rpm, 10 min). The water-soluble inositol phosphates were collected (upper phase) and separated by ion exchange on Dowex AG 1-X8 (chloride form). The used eluents were as follows: H 2 O (inositol), 5 mM sodium tetraborate, 60 mM sodium formate (glycerophosphoinositol); 0.1 M formic acid, 1 M ammonium formate (for total InsP). In parallel samples, the chloroform phase was dried and counted (total incorporation). The 3 H content of each fraction was determined by liquid scintillation counting as described previously (12).
Binding Assay-Cells (5 ϫ 10 6 /well) were washed three times with the assay buffer (PBS containing 0.1% bovine serum albumin (pH 7.4)) and then incubated for 60 min at room temperature with 125 I-labeled GnRH-A (100,000 cpm/ml) and increasing concentrations of unlabeled GnRH-A. Incubation was terminated by washing each well three times with PBS (pH 7.4). NaOH (1 N, 0.3 ml) was then added at room temperature, and after 60 min the cells were collected. Following the addition of 0.3 ml of 1 M Tris-HCl (pH 7.4), the 125 I counts were determined using gamma counter as described previously (13).
Immunoblotting of the GnRHR Mutants-Forty eight hours after transfection, membranes were prepared from COS-1 cells that were removed from dishes by scraping followed by homogenization in a Dounce homogenizer (30 strokes). The nuclei were then pelleted by centrifugation at 750 ϫ g for 15 min. The supernatant was collected and recentrifuged at 60,000 ϫ g for 30 min to obtain the membrane pellet. The pellet was collected and resuspended in sample buffer for protein separation on 10% SDS-PAGE, followed by Western blotting with mouse monoclonal antibody directed at the HA epitope tag (10). The blots were developed with alkaline phosphatase or horseradish peroxidase-conjugated anti-mouse or anti-rabbit Fab antibodies (Jackson ImmunoResearch).
Receptor Internalization-Forty eight hours after transfection, the cells were washed in ice-cold buffer I and then incubated with 100,000 cpm 125 I-GnRH-A for 3 h on ice. The cells were then moved to a 37°C water bath and incubated for the indicated times to allow internalization, without removing the radiolabeled peptide from the medium. After the incubation, the cells were transferred to an ice bath and washed twice with cold PBS. Externally bound ligand was collected by a 10-min acid wash (50 mmol/liter acetic acid, 150 mmol/liter NaCl), whereas the internalized ligand was measured by solubilizing the cells with 0.1 mol/liter NaOH as above.
Activation of Mitogen-activated Protein Kinase Cascades-Cells were grown in 6-well plates, serum-starved (0.5% FCS) for 16 h, and later stimulated with GnRH, and the cells were washed twice with ice-cold PBS and once with ice-cold buffer A (50 mM ␤-glycerophosphate, 1.5 mM EGTA, 1 mM EDTA, 1 mM dithiothreitol, 1 mM sodium orthovanadate). Cells were harvested in 0.3 ml of buffer H (buffer A containing 1 mM benzamidine, 10 g/ml aprotinin, 10 g/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride) followed by sonication (two times for 7 s, 40 watts) and centrifugation (15,000 ϫ g, 15 min, 4°C). The supernatants were collected, and aliquots (30 g/sample) were separated on 10% SDS-PAGE, followed by Western blotting with mouse monoclonal antiphospho-ERK. Total ERK was detected with a polyclonal antibody as a control for sample loading. The blots were developed with alkaline phosphatase or horseradish peroxidase-conjugated anti-mouse or antirabbit Fab antibodies (Jackson ImmunoResearch). The blots were autoradiographed on Kodak X-100 films, and the phosphorylation was quantitated by densitometry (Bio-Rad 690 densitometer).

GnRH-induced InsP Formation Is Attenuated by Prior
Exposure of the Cells to a GnRH Challenge-Pretreatment of ␣T3-1 cells with GnRH for 1 or 2 h abolished the ability of a second stimulus of GnRH to elevate InsP levels (Fig. 1A). To gain further insight into the kinetics of the effect, a kinetic response of GnRH-induced InsP elevation was performed. ␣T3-1 cells were treated with GnRH for varying times, and InsP levels were determined. As shown in Fig. 1B, the rate of elevation of InsP levels in the first 10 min was 10-fold higher than in the following 10 -60 min of incubation (1.1-and 0.1-fold/min, respectively). The results suggest that an inhibitory process at the receptor level, or at a more distal downstream signaling event, was initiated after 10 min of ligand stimulation. Based on these results we used GnRH elevation of InsP after 30 min of treatment as a marker for receptor activation in the rest of the experiments.
Synthetic Peptides Corresponding to ICL2 and ICL3 but Not ICL1 of the GnRHR Potentiate InsP Formation by GnRH-As mentioned before, one of the unique structural features of the mammalian type I GnRHR is the absence of a C-terminal tail. Therefore, to test the hypothesis that the ICLs may be involved in GnRHR regulation, synthetic peptides corresponding to the ICLs of the GnRHR were synthesized. Because ICL3 of the GnRHR is relatively long, we prepared separate peptides to the 13 N-terminal and the 13 C-terminal residues of ICL3. The sequences and abbreviations of the peptides are summarized in Table I. A concentration of saponin (50 g/ml) was found to achieve optimal cell permeabilization as described previously (11,14). Because permeabilization of the cells disrupts intracellular Ca 2ϩ levels, we first tested whether the extracellular Ca 2ϩ concentration affects the GnRH-induced InsP elevation in the permeabilized cells. The free Ca 2ϩ concentrations were calculated using the Calcon program. The maximal elevation in InsP production induced by GnRH was obtained in a free Ca 2ϩ concentration of 10 Ϫ9 M (not shown). Hence, this concentration of Ca 2ϩ was used in all experiments.
We then tested whether the synthetic peptides that correspond to ICL1, ICL2, or ICL3 affect GnRHR activation. Permeabilized ␣T3-1 cells were incubated with the peptides (each at 100 M). After 1 h of recovery, the cells were treated with or without GnRH for 30 min, and InsP levels were determined. None of the synthetic peptides altered basal InsP levels. The ICL2 peptide gave a marked enhancement of the GnRH response (p Ͻ 0.001; Fig. 2). The two peptides comprising the N and the C termini of ICL3 induced a smaller enhancement of GnRH-induced InsP elevation. In contrast, ICL1 had a small inhibitory effect. The data suggest that ICL-2 and to a lesser degree ICL3, but not ICL1, in the GnRHR harbor inhibitory sites. Therefore, the exogenously added ICL2 and ICL3 peptides may serve as "decoy peptides" and protect the receptor ICL inhibitory sites from being targeted by intracellular proteins involved in desensitization.
Preincubation of ␣T3-1 cells with increasing concentrations of the ICL peptides was then performed (Fig. 3). The peptides had no effect on basal InsP levels. The ICL1 peptide had no effect on GnRH-induced InsP elevation (Fig. 4A). However, the ICL2 peptide enhanced GnRH-stimulated InsP production in a dose-response manner (Fig. 3B). The two peptides ICL3N and ICL3C displayed a concentration-dependent potentiating effect on GnRH-induced InsP elevation (Fig. 3, C and D), albeit less pronounced as that obtained with the ICL2 peptide. Because the maximal effect was observed at a concentration of 100 M of the peptides, this concentration was used for the rest of the experiments. This is a relatively high concentration compared with GnRHR in cells, and probably reflects the high flexibility of small synthetic peptides and their limited ability to mimic the constrained loop structure in the receptor. We reasoned that if the peptides are titrating an inhibitor such as an RGS protein, then other G q -linked receptors are also likely to be affected. We therefore repeated the experiment with GnRH (as a positive control) and with other ligands known to interact with pituitary gonadotrophs, namely pituitary adenylate cyclase-activating polypeptide (PACAP), oxytocin, and endothelin (each at 100 nM). Only GnRH-induced InsP production was potentiated by the ICL2 peptide (not shown), indicating specific interaction of the inhibitory protein with the GnRHR.
As the ICL peptide effects on GnRH stimulation of InsP may have been due to changes in GnRH binding to its receptor, the synthetic peptides were inserted into permeabilized ␣T3-1 cells, and competitive binding analysis was performed. As shown in Fig. 4, the peptides had no significant effect on the binding characteristics of 125 I-GnRH analog (GnRH-A) to ␣T3-1 cells, except for a small elevation in the affinity of the receptor observed for the cells that were inserted with ICL3C (Fig. 4).
Ser 153 Plays a Key Role in the Protective Effect of ICL2-The putative importance of Ser and Thr residues in ICL2 of other GPCRs (15,16) led us to synthesize modified peptides, in which the Ser or Thr residues were changed to Ala, and to use these peptides to test whether one of the Ser/Thr residues is responsible for the stimulatory effect of the ICL2 peptide. Because the Ser and Thr residues are potential targets for various kinases, which may participate in desensitization, we also synthesized modified phosphomimetic peptides, in which the Ser or Thr residues were changed to Glu, which is predicted to mimic phosphorylation. The sequences and abbreviations of the peptides are given in Table I. ␣T3-1 cells were permeabilized and incubated with ICL2-WT, ICL2-Ala, or ICL2-Glu, in which one Thr and two Ser residues were replaced by Ala or Glu, respectively (Table I). Results in Fig. 5 show that as before the addition of the ICL2-WT peptide enhanced GnRH-induced InsP formation. However, the protective effect of the ICL2-WT peptide was abolished when the cells were incubated with the ICL2-Ala or the ICL2-Glu peptides. The results suggest that one or more of the Ser or Thr residues in ICL2 of the GnRHR are involved in the potentiating effect of ICL2 on the GnRH response, and this is phosphorylation-independent. Therefore, six peptides in which only one of the Ser or Thr residues was replaced with Ala or Glu were inserted into permeabilized ␣T3-1 cells, and GnRH-induced InsP elevation was determined (Fig. 6). As before, addition of ICL2-WT enhanced the GnRH response, and this potentiating effect was lost upon the insertion of ICL2-Ala or ICL2-Glu. Of the peptides with single substitutions, only ICL2-T144A and ICL2-T144E retained the potentiating effect of the ICL2-WT. The peptides ICL2-S151A and ICL2-S151E enhanced the GnRH response, albeit to a lesser degree than the ICL2-WT. Only replacement of Ser 153 to Ala or Glu (ICL2-S153A and ICL2-S153E) restored the inhibitory effect of the ICL2-Ala and ICL2-Glu peptides. The observation that the Ala and the Glu mutants of ICL2 gave similar results suggests that the Ser or the Thr residues of ICL2 are not phosphorylated during receptor activation and are responsible for desensitizing the receptor by a phosphorylation-independent mechanism. Indeed, Ser 151 and Ser 153 are not known as potential phosphorylation sites in the ICL2 of the GnRHR (17). The data point to an alternative, possibly novel role for the Ser/Thr residues of the ICL2 of type I GnRHR as a core of an inhibitory domain in which Ser 153 and to a lesser degree Ser 151 are key residues.

GnRH Does Not Induce Receptor Phosphorylation in ␣T3-1
Cells-To corroborate this proposal further, we examined whether the GnRHR undergoes phosphorylation after a GnRH challenge. After overnight serum starvation, ␣T3-1 cells were labeled for 4 h at 37°C in P i -free DMEM containing 150 Ci/ml 32 P i and were washed. GnRH was added for various times, and cell extracts were resolved by SDS-PAGE and visualized using a PhosphorImager as described for the AT1 angiotensin receptor (18). GnRH has no consistent and reproducible effect on receptor phosphorylation. 2 Our data are in agreement with Willars et al. (19).
Characterization of the Mutated GnRH Receptors in COS-1 Cells-To analyze further the role of ICL2 and specifically the role of Ser 153 in GnRHR activation, we prepared two GnRHR mutants. In one of the mutants (ICL2-Ala), two Ser residues (Ser 151 and Ser 153 ) and one Thr (Thr 142 ) were mutated to Ala, and in the other mutant only Ser 153 was mutated to Ala (S153A). In preliminary studies, we checked whether we could obtain similar binding kinetics in transient transfection of COS-1 cells with the WT GnRHR as compared with ␣T3-1 cells. The results showed similar K d values for the ␣T3-1, which expresses the native receptor, and the GnRHR-transfected COS-1 cells (K d 0.14 Ϯ 0.05 and 0.23 Ϯ 0.04 nM, respectively). A negative control consisting of cells transfected with the vector alone did not exhibit any binding (not shown). We then examined the binding properties of the WT GnRHR and the two mutants ICL2-Ala (S151A, S153A, and T142A) and S153A in the transfected COS-1 cells. Binding competition studies with 125 I-GnRH-A revealed unchanged affinity between the WT and the mutant GnRHRs ( Fig. 7A and Table II). Similarly, no significant differences were observed in the expression levels of the HA-FLAG WT and the mutated GnRHR in membranes from transfected COS-1 cells (not shown). To rule out the possibility that receptor activation can differ because of differences in the rate of receptor internalization, cells were transfected with WT or mutated GnRHRs, incubated with 125 I-GnRH-A on ice, and then transferred to 37°C to allow internalization. An acid wash was used to remove cell surface-bound ligand, and internalized ligand was measured after solubilizing the cells. As seen in Fig. 7B, agonist-induced internalization was similar for both the WT and the mutated receptors.
Based on these results, COS-1 cells were transfected with the WT and the mutants of the GnRHR receptor (ICL2-Ala and S153A), and the time course of GnRH-induced InsP production 2 R. D. Smith, Z. Naor, and K. J. Catt, unpublished observations.  Table I   was determined. As shown in Fig. 8A, addition of GnRH to the WT-transfected cells resulted in a 5-fold increase in InsP formation, which reached a peak at 60 min with a t1 ⁄2 of 16 min. On the other hand, stimulation of GnRH in the ICL2-Ala-and S153A-transfected cells also resulted in a 5-fold increase in InsP formation. The peak response was reached already at 30 min, and a remarkable reduction of t1 ⁄2 to 2.5 min was found. Hence, the rate of GnRH-induced elevation of InsP formation was 6.5-fold higher, particularly in the first 10 min in the cells that were transfected with the mutated receptors as compared with the cells with the WT receptor. However, WT and the mutated receptor had similar maximal InsP levels. To assess the role of Ser 153 in GnRH-induced ERK1/2 activation, COS-1 cells were transfected with the WT and the mutants of the GnRHR receptor (ICL2-Ala and S153A), and ERK1/2 activation was determined (Fig. 8B). Both mutants enhanced ERK1/2 activation by GnRH. The results support our proposal that the GnRHR harbors inhibitory domains within the ICL2 and that Ser 153 plays a key role in this domain.

DISCUSSION
Signal termination, receptor desensitization and re-sensitization, and down-regulation are regulated processes mediated by covalent modifications, association with intracellular proteins, internalization, and trafficking of activated GPCRs (1, 2). Rapid homologous desensitization often involves partial or complete uncoupling of the receptors from the effector proteins, which may occur within seconds to minutes of agonist occupancy (1,2). Considerable evidence has implicated ICL2 and ICL3, as well as the membrane-proximal region of the C terminus of several GPCRs, as involved in G-protein coupling and determination of signal specificity (20,21). Phosphorylation of GPCR by serine/threonine protein kinases, predominantly through GRKs on phosphorylation sites localized in the Cterminal tail and ICL3, facilitates the binding of arrestin to intracellular domains of GPCRs (1,2,22). Arrestin binding induces uncoupling from the G-proteins and facilitates receptor internalization via components of the clathrin endocytic apparatus culminating in receptors being recycled to the cell surface or proteolytically degraded in lysosomes (1,2,5,23,24).
Another level of receptor regulation is maintained by members of the regulators of G-protein signaling (RGS), a large family of proteins that modulate G-protein activity. RGS proteins interact directly with active G␣ subunits to accelerate their intrinsic GTPase activity and to limit their half-life, hence curtailing or terminating their activity (25). It was shown recently that RGS2 binds directly to the M1 muscarinic acetylcholine receptor ICL3 to modulate G q/11 ␣ signaling (26). Two family members, RGS3 and RGS10, have been implicated in the regulation of GnRH receptor coupling (27)(28)(29), and the C-terminal tails of nonmammalian GnRH receptors may be sites for interactions with RGS10, although the nature of this interaction is unclear (27). Hence, C-terminal tail phosphorylation and modulation by RGSs are now regarded as the main mechanisms leading to desensitization and/or signal termination for GPCRs.
In addition there is precedence for involvement of accessory proteins such as arrestin, GRKs, Src homology 2 domain-containing proteins, small GTP-binding proteins, polyprolinebinding proteins, receptor activity-modifying proteins, and members of the scaffolding family of proteins such as PDZ domain-containing proteins in the regulation of GPCR signaling in general. However, the requisite sequence structural motifs in the GnRHR responsible for such interactions are largely unknown (17).
We noticed that the rate of production of InsP stimulated by GnRH in the first 10 min was 10-fold higher than in the following 10 -60 min of incubation (1.1-and 0.1-fold/min, respectively). This could result potentially from a receptor phosphorylation that was followed by desensitization. Alternatively, binding of an inhibitory accessory binding partner to the GnRHR could also have triggered the decline in receptor activity. We have therefore utilized synthetic peptides corresponding to the ICLs of the GnRHR to shed light on the mechanisms of signal termination. Synthetic peptides have been shown to modulate receptor and G-protein activities in numerous systems, including the rhodopsin, ␤and ␣2-adrenergic, muscarinic, and dopamine D 2 receptors (8,30,31). Hence, such peptides may be used to further our understanding of the structure-function relationship of the GnRHR. Although synthetic peptides are usually used in cell-free systems (9), we decided to use cell permeabilization for delivery of the peptides (14). The synthetic peptides had no effect on the basal InsP level, suggesting that activation of G q by the GnRHR (12, 32) involves structural determinants from more than one ICL. Indeed, G q activation by the agonist-occupied GnRHR is GnRHRs. Cells were transfected as above, and 48 h after the transfection the cells were serum-starved for 8 h and then subjected to 10 min of treatment with GnRH (100 nM). ERK1/2 activity was determined using specific antibodies to phosphorylated ERK. The experiment was performed in duplicate, and the immunoblot is a representative of two different experiments. thought to involve determinants in ICL2 (Pro 146 and Leu 147 ) and ICL3 (Arg 240 , Val 241 , Leu 242 , Arg 260 , Ala 261 , and Arg 262 ) (17). Although the peptides had no effect on basal InsP formation, the ICL2 peptide of the GnRHR potentiated GnRH-induced InsP production. A smaller effect was also observed for the ICL3-derived peptides. Because ICL1 and the "mutated" analogs of ICL2 had no effect, the results indicate that the effect exerted by the other peptides was specific. Binding studies showed that introduction of the peptides into ␣T3-1 cells had no significant effect on GnRH binding.
Conventional GPCR uncoupling and desensitization is mediated through GRK and/or protein kinase A and PKC phosphorylation of the C-terminal tail and ICL3. For GnRHR, we contemplated that the ICLs substitute for the C-terminal tail for this mechanism. If this were the case, the synthetic ICL peptides would act as alternative substrates and protect receptor ICLs from phosphorylation (i.e. as a decoy). However, as mentioned above, mammalian type I GnRHR does not undergo rapid agonist-dependent phosphorylation and desensitization (5), making it unlikely that the peptides would protect the receptor from phosphorylation-dependent desensitization. Indeed, previous studies (19) have shown that the GnRHR is not phosphorylated upon agonist stimulation; a finding confirmed in our current studies. Moreover, GnRHR internalization is independent of ␤-arrestin (5). It therefore appears that the peptides are alternative targets for other intracellular proteins, which bind to and modulate G␣ q coupling to GnRHR. Potential candidates are RGSs, Src homology 2 domain-containing proteins, polyproline-binding proteins, receptor activity-modifying proteins, and members of the PDZ domain proteins, which are all known to regulate GPCR signaling (17). Therefore, the exogenously added ICL peptides most likely have served as decoy peptides and protected against interaction of the inhibitory proteins with ICL receptor sites.
We therefore suggest that the synthetic peptides mimic receptor ICL domains that are targets for inhibitory proteins, and thus enable further studies to characterize those domains and the accessory proteins that form a complex during receptor activation leading to signal termination. If this is correct, residues of ICL2 and ICL3 are involved in the regulation of receptor activation. Similar results were obtained for the luteinizing hormone receptor, where a synthetic peptide corresponding to the entire ICL3 reversed desensitization of adenylyl cyclase activity (33).
Because introduction of ICL2 into permeabilized ␣T3-1 cells had the most significant potentiating effect on GnRH-induced InsP production, we decided to investigate the residues responsible for the effect. Ser/Thr phosphorylation of GPCRs is one of the most prevalent events in receptor regulation. Three Ser residues (Ser 140 , Ser 151 , and Ser 153 ) and one Thr residue (Thr 144 ) are present in ICL2, suggesting that phosphorylation of these residues may be responsible for the ICL2 effect. Hence, to elucidate the role of the Ser/Thr residues of ICL2 in GnRHR regulation, we synthesized mutated peptides in which all the Ser and Thr residues of ICL2, or only one of them, were mutated to Ala or Glu. The stimulatory effect of ICL2 on GnRHinduced InsP formation was abolished when the peptides, in which all the Ser or Thr residues were mutated to Ala or Glu (ICL2-Ala, ICL2-Glu), were inserted to the permeabilized ␣T3-1 cells, implicating Ser and Thr residues of ICL2 in the protective effect. When the peptides with single substitution of Ser/Thr were inserted to the cells, the two peptides in which Ser 153 was replaced by Ala or Glu (ICL2-S153A and ICL2-S153E, respectively) did not show the increase in GnRH-induced InsP elevation observed for the ICL2-WT peptide. These results indicate that Ser 153 , and to a lesser extent Ser 151 , are important residues in the protective effect of the ICL2 peptide. Therefore, the peptides, in which Ser 153 was replaced by Ala or Glu, failed to serve as decoy peptides. We had expected that the role of Ser 153 be as a phosphorylation target, which results in receptor desensitization. The failure of the Ala 153 to enhance GnRH-stimulated InsP production supports this notion. However, the phosphomimetic Glu 153 peptide would then be expected to compete effectively for desensitizing proteins that bind to phosphorylated Ser 153 and enhance stimulation of InsP production as well as, or better than, the WT receptor. Thus, the failure Glu 153 to enhance GnRH-stimulated InsP production suggests that the intracellular proteins involved in desensitization of GnRHR must bind to a domain that involves Ser 153 or is configured by Ser 153 in an unphosphorylated state. This is supported by Willars et al. (19), and our demonstration that GnRHR is not phosphorylated by GnRH.
To confirm further the role of ICL2, and in particular that of Ser 153 , in GnRHR activation and signal termination, we transfected COS-1 cells with the WT and two mutants, ICL2-Ala (in which Ser 151 , Ser 153 , and Thr 142 were mutated to Ala) and S153A, and followed agonist-dependent receptor activation. First, a good agreement was noticed between the K d values for GnRH binding in the ␣T3-1, which expresses the native receptor, and the GnRHR-expressing COS-1 cells. Second, binding studies with 125 I-GnRH revealed similar parameters between WT and the mutant GnRHRs expressed in the COS-1 cells. Also, no significant differences were observed in the expression levels of the HA-FLAG WT and the mutated GnRHR in membranes from transfected COS-1 cells, and agonist-induced internalization of the GnRHR was similar for both the WT and the mutated receptors.
GnRH stimulated a 5-fold increase in InsP production in WT GnRHR-expressing COS-1 cells, which reached a peak at 60 min with a t1 ⁄2 of 16 min. Although a 5-fold increase in InsP production was also observed in the two mutated GnRHRs, the peak response however was already reached at 30 min, and a 6.5-fold reduction of t1 ⁄2 to 2.5 min was observed. In addition, ERK activation by GnRH was enhanced in the two mutated GnRHRs. The results support the notion that the GnRHR harbors inhibitory domains within the ICL2 and that Ser 153 plays a key role in this domain. Nevertheless, a basic difference was observed between the results obtained with the exogenously added peptides of the ICLs in ␣T3-1 cells and the transfection studies with the WT and the mutants in COS-1 cells. Although in ␣T3-1 cells the peptides enhanced the maximal response of GnRH-induced InsP formation at 30 min, the mutants had no such effect in the transfected COS-1 cells but had a remarkable effect on the rate of the response. We therefore assume that the inhibitory domains in ICL2 and Ser 153 manifest themselves differently in different cell type backgrounds because of differences in the intracellular protein milieu.
The ICL2 of the GnRHR was reported in several studies to be involved in signal propagation and G-protein selectivity (17). Although mutations in Ser 140 resulted in an impaired internalization process (34), mutation of Arg 139 (part of the DRY motif) to Gln significantly reduced InsP production but did not affect internalization (35). Mutation of another conserved residue, Leu 147 , to Ala or Asp, also impaired InsP formation (34). Coexpression of a WT GnRHR with a truncated form of the GnRHR, which lacks one-third of the C-terminal region, including ICL3, significantly impaired the signaling ability of the receptor, probably due to interactions of the wild type receptor with the truncated form (36). Exploration of the structural characteristics of ICL2 of the GnRHR using the computational method of conformational memories showed that the wild type ICL2 loop has accessible states that can interact with ICL3 (37). However, mutation of the conserved sequence TRPLA of ICL2 to a more constrained TPPLA sequence prevented most of the conformational states of the Pro-Pro mutant from interacting with ICL3. Mutagenesis of Arg in the TRPLA motif to Pro markedly reduced the receptor efficiency, suggesting that ICL2-ICL3 interaction is necessary for efficient G-protein coupling. Collectively, the above studies show that ICL2 is necessary for proper coupling of Gn-RHR to G q . We therefore propose that binding of ICL2 to an inhibitory accessory protein, such as an RGS family member, results in the disruption of GnRHR-G q interaction and signal termination. Our results identify Ser 153 as a core in this binding pocket, and we are currently exploring a proteomics approach to identify putative ICL2-binding proteins using WT and mutated ICL2 domains that may be involved in signal termination.